Color profile: Disabled Composite 150 lpi at 45 degrees Acta Bot. Croat. 71 (1), 87–94, 2012 CODEN: ABCRA25 ISSN 0365–0588 eISSN 1847-8476 DOI: 10.2478/v10184-011-0055-4 The correct phylogenetic position of Lotus conimbricensis Brot. (Leguminosae, Loteae) based on nuclear ribosomal ITS sequences MIGUEL A. FARIA1,D.JAMES HARRIS2,3,*,TATIANA VISNEVSCHI-NECRASOV2, MANUEL TAVARES DE SOUSA4,EUGÉNIA NUNES2 1 REQUIMTE, Laboratório de Bromatologia e Hidrologia, Faculdade de Farmácia da Universidade do Porto, Rua de Aníbal Cunha, 164, 4099-030 Porto, Portugal 2 CIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, Campus Agrário de Vairão, 4485-661 Vairão, Portugal. 3 Departamento de Biologia, Faculdade de Ciências da Universidade do Porto, 4099-002 Porto, Portugal. 4 Estação Nacional de Melhoramento de Plantas, Estrada de Gil Vaz, 7350-951 Elvas, Portugal Abstract – The nrDNA ITS sequence determined in Lotus conimbricensis in a previous phylogenetic study was unusual, in that it was almost identical to those retrieved from the morphologically distinct species L. subbiflorus. In the present study we sequenced new specimens of both species to reassess the phylogenetic position of L. conimbricensis.We conclude that the ITS sequence of L. conimbricensis used in the earlier analyses was most likely erroneous, and in fact L. conimbricensis is not closely related to L. subbiflorus. Crit- ical reexamination of previously published data indicates that several other similar errors may exist for other Lotus species, and these should be checked before taxonomic conclu- sions are made. Key words: Lotus conimbricensis, Lotus subbiflorus, ITS, phylogeny, taxonomy, nrDNA Introduction Lotus is the largest genus within the tribe Loteae, with approximately 130 species. His- torically there has been little agreement in the taxonomic literature regarding the generic limits of Lotus and its infrageneric subdivision (DEGTJAREVA et al. 2006). However, this has changed considerably with the advent of phylogenetic studies based on nrITS sequences. These have clearly shown that the New World species of Lotus are not closely related to the * Corresponding author, e-mail: [email protected] Copyright® 2012 by Acta Botanica Croatica, the Faculty of Science, University of Zagreb. All rights reserved. ACTA BOT. CROAT. 71 (1), 2012 87 U:\ACTA BOTANICA\Acta-Botan 1-12\481 Faria.vp 26. o ujak 2012 11:00:29 Color profile: Disabled Composite 150 lpi at 45 degrees FARIA M. A., HARRIS D. J., VISNEVSCHI-NECRASOV T., TAVARES DE SOUSA M., NUNES E. Old World species (ALLAN and PORTER 2000), and in particular DEGTJAREVA et al. (2006) revised sectional classifications proposed by KRAMINA and SOKOLOFF (2003) and SOKOLOFF (1999a, b). Some sections appeared as non-monophyletic, including the section Lotus, which was resolved as paraphyletic since Lotus conimbricensis Brot. (Lotus sect. Erythrolotus Brand) had an ITS sequence type identical to those found in Lotus subbiflorus Lag. (Lotus sect. Lotus). Most Lotus species are distinct from each other in their ITS sequences, so it is surpris- ing that two morphologically dissimilar species such as L. subbiflorus and L. conimbri- censis should be identical in this DNA region. Errors in GenBank are well known (HARRIS 2003), and so unusual results that may have important taxonomic implications deserve careful investigation. Furthermore, the ITS region is known to display intra-individual variation in some taxonomic groups, that can confound phylogenetic studies (HARRIS and CRANDALL, 2000). DEGTJAREVA et al. (2006) combined new data with previously published sequences and noted that in the case of Lotus creticus they obtained a very different se- quence and phylogenetic placement when compared to the sequence by ALLAN et al. (2003, 2004). The aim of this study was to sequence several individuals of both L. subbiflorus and L. conimbricensis to resolve their phylogenetic relationship, and to look for some other possi- ble examples of discordance in previous studies by examination of sequence data available for different Lotus species. Materials and methods Specimens were obtained from the Estação Nacional de Melhoramento de Plantas (Elvas, Portugal). DNA was isolated from leaf tissue using standard methodologies (SAM- BROOK et al. 1989). The entire ITS1 and 1TS2 region was amplified using universal primers (WHITE et al. 1990). Amplifications were performed in a Biometra T3 thermalcycler (Bio- metra, Goettingen, Germany) in 20 mL reactions consisting of approximately 10 ng DNA template, 1 mM of each primer, 200 mM of each dNTP, 0.5 U EcoTAQ DNA polymerase (Ecogen, Barcelona, Spain), 2 mL of 10X PCR buffer and 1.5 mM MgCl2, using the follow- ing amplification protocol: initial denaturation at 95 °C for 2 min followed by 30 cycles of 95 °C for 30s, 53 °C for 30s and 72 °C for 1 min. A final extension step at 72 °C for 7 min- utes was performed. PCR products were purified using the JetQuick (Genomed, Löhne, Germany) micro spin kit based on a surface modified silica membrane and sequenced using the same primers on an ABI 3730 DNA sequencer (Applied Biosystems, Foster City, USA) using the kit BigDyeTerminator v3.1 from the same supplier. Six specimens of L. conimbricensis,andtwoL. subbiflorus were sequenced. 101 se- quences of Lotus were taken from GenBank, as well as the three closest outgroups, follow- ing DEGTJAREVA et al. (2006) – Cytisopsis pseudocytisus, Hammatolobium lotoides and Anthyllis tetraphylla. Sequences were aligned using ClustalW with default parameters (THOMPSON et al. 1997). Phylogenetic analysis was run using MrBayes v3.1 (HUELSENBECK and RONQUIST 2001), using the GTR+I+G model, as selected by Modeltest (POSADA and CRANDALL 1998). Parameters were estimated as part of the analysis, with four Markov chains. The analysis was run for 106 generations, saving one tree every 100 generations. 88 ACTA BOT. CROAT. 71 (1), 2012 U:\ACTA BOTANICA\Acta-Botan 1-12\481 Faria.vp 26. o ujak 2012 11:00:29 Color profile: Disabled Composite 150 lpi at 45 degrees THE PHYLOGENETIC POSITION OF LOTUS CONIMBRICENSIS The log-likelihood values of the sample point were plotted against the generation time and trees obtained prior to reaching stationary (25%) were discarded. Remaining trees were combined in a 50% majority consensus tree . Two independent runs were made to check for convergence. Results In total 106 taxa were analysed, including the three outgroup taxa. Aligned sequences were 646 nucleotides long (383 constant sites, 208 informative sites). As expected the two species sequenced as part of this study showed limited intraspecific variation – both sam- ples of L. subbiflorus were identical, while of the six samples of L. conimbricensis five were identical and one differed by a single C-G mutation. Therefore in the analysis only one sequence per species was included. All new sequence data have been deposited in GenBank (accession numbers JQ655098 to JQ655105). The phylogram estimated from the Bayesian analysis is shown in figure 1. The ITS sequences of L. subbiflorus generated here are identical to several other L. subbiflorus sequences from GenBank. One L. subbiflorus, however, has a very different sequence (AF450160), being identical to that from two speci- mens of Lotus arenarius (AF450193 and AF218528). The ITS sequence of L. conimbri- censis previously published (AF450186) is almost identical to L. subbiflorus as sequenced in this and other studies. However, sequences from our six samples of L. conimbricensis are very different from the previously reported sequence, and appeared in a different part of the tree (Fig. 1). The closest relative of L. conimbricensis would be a sample of Lotus halophilus (AF450208), which differs by just two or three mutations. However, another sample of Lotus halophilus (DQ160283) is extremely distinct and part of another clade. Discussion The use of nrITS sequences has revolutionized Lotus systematics. Nevertheless, some conclusions based on poor data or uncritical data treatment can be premature or even incor- rect. An example is the single specimen of Lotus conimbricensis, previously sequenced and found to be almost identical to Lotus subbiflorus. Later studies included the same sequence from GenBank and reached the same conclusions (e.g. DEGTJAREVA et al. 2006). However, it is now clear that several published sequences in GenBank are anomalous, with extremely divergent sequence types recovered by different authors from the same species. There may be various explanations for it. One is that in some groups intraindividual variation is very common (HARRIS and CRANDALL 2000). In these cases divergent copies obtained from the same individual can appear in different phylogenetic positions. This seems unlikely in the case of Lotus however, since when this happens many heterozygous positions are usually observed, unless the initial PCR products are cloned. Also, when multiple individuals are sequenced, variation would be expected. Although we sequenced six L. conimbricensis, only a single nucleotide difference was found. Another explanation may be that consider- able variation exists within species. For instance, DEGTJAREVA et al. (2006) identified an ITS sequence in L. creticus, which was very different from those reported by ALLAN et al. (2003), and suggested that further studies are needed to assess intraspecific variation within this species. Again, however, we think that this explanation is unlikely in most cases ACTA BOT. CROAT. 71 (1), 2012 89 U:\ACTA BOTANICA\Acta-Botan 1-12\481 Faria.vp 26. o ujak 2012 11:00:29 Color profile: Disabled Composite 150 lpi at 45 degrees FARIA M. A., HARRIS D. J., VISNEVSCHI-NECRASOV T., TAVARES DE SOUSA M., NUNES E. 90 ACTA BOT. CROAT. 71 (1), 2012 U:\ACTA BOTANICA\Acta-Botan 1-12\481 Faria.vp 26. o ujak 2012 11:00:31 Color profile: Disabled Composite 150 lpi at 45 degrees THE PHYLOGENETIC POSITION OF LOTUS CONIMBRICENSIS species estimated using a Bayesian approach. * Indicate nodes with >95% Bayesian Lotus Phylogram of relationships of available Posterior Probability support. Arrows indicate newly sequenced haplotypes. Fig 1. ACTA BOT.